It may be surprising to learn that the human cerebral cortex has more than 50 distinct areas that are devoted to the analysis of visual signals. Why so many areas needed to interpret the retinal image? One answer to this question is that vision is, by its nature, an interpretive process -- a much more difficult problem than one might think. Consider a few moments in the life of a human or any other primate. As we walk or climb through our environment, we turn our heads and move our eyes. We glance three times per second from object to object, in some cases trying to understand the expression on a face, and in others deciding how to shape our hand in order to establish a correct grip. Our own movements cause our retinal image to flit about in a manner that would be impossible to understand if we were to see it on a computer monitor. Nonetheless, our brain is able to interpret this incomprehensible sequence of retinal images and integrate it into a stable visual world: Indeed, the retina is only our visual sensor, controlled by the brain to sample data from our environment. It makes no sense to take each snapshot at face value instead the sequence of retinal images is, from the very beginning, interpreted within the context of the movements issued by the brain. Likewise, the brain makes an educated guess on the identity and use of objects, the distance to locations within the scene, the emotions and intentions of people, and a countless number of other features of the world. These qualities are not written into the visual signals themselves, but must be judged based on physical cues, experience, and intuition. In some cases the judgments are unconscious and automatic, whereas in others they are at the fore of our thought processes. In the this we focus on aspects of visual perception that are so immediate and intuitive to us that it is not at all obvious that there is a problem to be solved. In one subproject, we have been investigating how the brain infers a surface based on contextual cues. Specifically, we investigated the link between subjective surface completion and neuronal activity in cortical area V4 of two trained monkeys. Multi-electrode arrays measured single-unit responses to two categories of stimuli: Kanizsa stimuli gave rise to the perception of an illusory surface, whereas control stimuli, with the inducing wedges rotated, did not. Many V4 neurons exhibited a markedly greater, and sometimes rhythmic, response to the Kanizsa configurations compared to controls. Importantly, enhanced firing depended on the precise position of receptive field focus (RF-focus), the spatial location of the peak stimulus sensitivity assessed by independent mapping, relative to the illusory surface. Neurons whose RF-focus fell directly on the surface showed the maximum response enhancement during subjective surface completion. Most neurons whose RF-focus fell upon an inducer failed to show such enhancement. These results demonstrate that although V4 neurons have sizeable receptive fields, their contribution to perceptual organization operates at much higher spatial acuity. In another subproject, we are investigating the role of the visual thalamus in several aspects of visual perceptual processing. This research follows directly from our previously published studies in which we measured neural correlates of perceptual suppression in the pulvinar and lateral geniculate nucleus (LGN), and investigated the role of the LGN in the phenomenon of blindsight. In the present study, we aim to map the complex pulvinar nuclei, along with the LGN, with respect to several cognitive parameters. The first such test involves activity during the phenomenon of binocular rivalry, in which nonmatching monocular stimuli presented to the two eyes result in a sequence of perceptual reversals perception is first dominated by the left-eye stimulus, followed a few seconds later by the right-eye stimulus, etc. With this phenomenon, the physical stimulus is always constant, but the perceptual experience varies. To what extent does activity in different segments of the pulvinar reflect the ever-changing perceptual state of a nonhuman primate? In order to address this question, we have trained three monkeys to report their perceptual experience using eye movements. We will soon be able to describe functional difference between pulvinar subdivisions with respect to their role in perceptual switching. The second test involves visual attention. For this test, we plan to chart the modulatory effect of visual attention throughout the pulvinar. The third test involves responses to complex, natural stimuli. In all cases, we predict that the types of responses we observe in different pulvinar segments will show a functional similarity to the response properties of the cortical regions known to project there. The results will have broad implications not only for the understanding visual perception, but also for the functional anatomy of the pulvinar, and for the nature of thalamocortical interplay more generally. In a final subproject, we have conducted a follow-up study to our blindsight experiments from the previous years. In this investigation we ask to what extent does visual information reach neurons in higher cortical areas when transmission via the primary visual cortex (V1) is absent? Previous reports addressing this question at the level of area V4 reached opposite conclusions, possibly due to methodological differences: Temporary cooling of macaque V1 resulted in the virtually complete elimination of neuronal spiking in V4 (Girard et al. Neuroreport, 1991). In contrast, fMRI measures of V4 in monkeys and human patients with permanent V1 injury revealed weak, but reliable modulations to visual stimulation in the affected part of the visual field (scotoma) (Goebel et al., Vision Research, 2001; Schmid et al., Nature, 2010). In our current experiments, we used chronically implanted electrode arrays to longitudinally record the neural activity in area V4 of 2 macaque monkeys before and after removing part of V1 by surgical aspiration. Care was taken to ablate the part of V1 that spans the horizontal meridian between 2-7 degrees of visual eccentricities (scotoma region) while leaving the lower vertical meridian representation intact (control region). Neural responses collected during the first month post-lesioning were evaluated by measuring multi-unit spike modulation during various visual stimulation conditions (reverse correlation receptive field mapping, spatially restricted grating patterns, full-field movie scenes). As expected, lesioning V1 had a devastating effect on the capacity of most V4 sites to respond to visual stimulation. Nevertheless, a minority of sites with receptive fields covering the visual field around the horizontal meridian showed significant multi-unit responses to visual stimuli presented to the scotoma region. However, in these cases modulation amplitudes reached only levels of 10 % compared to pre-lesioning. Interestingly, stimuli placed next to the scotoma boundary generated sometimes responses of higher magnitude compared to pre-lesion conditions, possibly indicating a release of horizontal inhibition underlying neural reorganization. Taken together, our results are in line with the view of the primary visual cortex as the major driver for neural activity in higher cortical areas. At the same time, the presence of weak responses to visual stimulation in the scotoma region supports the notion of V1-bypassing thalamic projections systems as alternative relays for the transmission of information to visual association cortex.